Health status measures can play an important role in evaluating the efficacy and effectiveness of pharmaceutical therapies. Health status measures may be especially important when outcomes such as death are infrequent, and the most important difference between treatments are in terms of quality of life. Health status measures are beginning to be incorporated into AIDS clinical trials. However, there are many unresolved issues in the use of these measures. Few studies have examined the relationship of health status to physiologic measures, or investigated their responsiveness to clinical change. Still fewer have described the effects of acute illness and its treatment on health status. Understanding is limited about how these measures relate to, differ from, and supplement conventional clinical measures. This methodologic study proposes to examine the reliability and validity of a brief health status measure for acute Pneumocystis carinii pneumonia (PCP). The study will analyze data being collected in a randomized AIDS clinical trial comparing 3 treatment regimens for acute PCP. The primary goal is to test the responsiveness of health status scales to physiologic changes (in terms of alveolar-arterial gradient) and other clinical changes over time. The study also aims to test the reliability and construct validity of the scale, and to examine the cross-sectional relationship of the health status scales to physiologic measures at several points in time. In addition, the study will explore what the measures add to conventional assessments, and compare the equivalence of a Spanish translation of the instrument. The study will be among the first to evaluate a health status measure for acute illness, and the first study of a health status measure in an AIDS clinical trial for acute infection. The results will be useful for planning future AIDS trials incorporating health status measures, especially for pulmonary disease. The approach should also be applicable to other studies of acute illness superimposed on chronic disease. In addition, the study will provide insights about the added value provided by health status measures, and about whether a translated measure is equivalent enough to allow data to be aggregated across linguistically diverse populations.